franck cassez (@franckDownunder) / X

franck cassez

590 posts

franck cassez

@franckDownunder

Head of research @movementlabsxyz

Sydney, New South Wales

Joined November 2014

franck cassez
@franckDownunder
Apr 1, 2025
Excited to be visiting @NTHU_TAIWAN this week for research collaborations! Looking forward to insightful discussions on Move, compiler optimisations, cryptography & formal verification, and blockchain security. #Blockchain #MoveLang #FormalVerification
17K
franck cassez
@franckDownunder
Feb 19, 2025
Replying to @SashaSpiegelman
Thanks @SashaSpiegelman for confirming what I thought about Monad tech. I was just reading the Monad docs this morning and could not find any mention of BlockSTM … although they mention STM! And speculative /optimistic execution with retry is clearly… blockSTM. BlockSTM is a
3.3K
franck cassez
@franckDownunder
Oct 26, 2021
We just released an experience report and code for our project: Verification of the #BeaconChain in Dafny. Was great fun and thanks @adiasg and the Ethereum Foundation for your support. #Ethereum #verification cs.paperswithcode.com/paper/formal-v…
franck cassez
@franckDownunder
Nov 25, 2020
The Automated Verification team in ConsenSys Software R&D have developed the first fully mechanised correctness proof of the Incremental Merkle tree algorithm. The code is written in Dafny, a verification-friendly programming language. github.com/ConsenSys/depo… #ETH2
GitHub - Consensys/deposit-sc-dafny: Deposit smart contract in Dafny
From github.com
franck cassez
@franckDownunder
Nov 25, 2024
Replying to @apenzk @toghrulmaharram and @rushimanche
@toghrulmaharram I don't know why the discussion is so tense, but here is my 2 cents: 1/7 (L1) pre-confirmation (based sequencing) = scheduled for execution L2-confirmation = executed + attested by >= 2/3 (super-majority) of L2 stake
13K
franck cassez
@franckDownunder
Mar 16, 2025
Replying to @hrkrshnn @lightclients and 2 others
@hrkrshnn "... rely on heuristics ...". That may not be completely true. I have developed a tool to generate CFGs from EVM bytecode (github.com/franck44/evm-d…) and it does not rely on heuristics. It uses abstract interpretation to generate a CFG and has some nice features: 1. it
GitHub - franck44/evm-dis: An EVM bytecode disassembler/assembler
From github.com
388
franck cassez
@franckDownunder
Nov 15, 2022
Want to verify EVM bytecode? We have developed a formal and executable semantics of the Ethereum Virtual Machine (EVM) in #dafny. The full power of verification-friendly languages to develop machine-checkable proofs of EVM bytecode. github.com/ConsenSys/evm-…. #formalverification
GitHub - Consensys/evm-dafny: An EVM interpreter in Dafny
From github.com
franck cassez
@franckDownunder
Apr 9, 2020
Leveraging state-of-the-art #formalmethods to verify the #ETH2 specs using #Dafny. github.com/PegaSysEng/eth… #PegaSysEng #Ethereum
GitHub - Consensys/eth2.0-dafny: Eth2.0 spec in Dafny
From github.com
franck cassez
@franckDownunder
Feb 19, 2025
Replying to @0xtrojan_ and @SashaSpiegelman
Why not? BlockSTM is an algorithm and was invented by them and others and implemented by them. They showed some scientific leadership 😀 and innovation. I don’t really understand what you mean by broad scientific concept.
202
franck cassez
@franckDownunder
Mar 21, 2025
that's a great podcast with clear and concise explanations of the tech and the spot-on comments and questions from the moderators.
Alexander Spiegelman
@SashaSpiegelman
Mar 21, 2025
Thanks @Tim_Roughgarden and @rhackett for having me, it was a blast! The only thing is, we ship sonfast at Aptos—by the time the episode was edited, we had already upgraded most of the tech we discussed 😅
319
franck cassez
@franckDownunder
Jan 28, 2025
Replying to @bfarmer
it is not only the zk-field, but Polygon implemented a parallel EVM 3 years ago (adapted the BlockSTM algorithm from Diem/Libra/Move).
Innovating the Main Chain: a Polygon PoS Study in Parallelization
From polygon.technology
299
franck cassez
@franckDownunder
Nov 25, 2024
Replying to @franckDownunder @apenzk and 2 others
2/7 Pre-conf is O(sec) (~100ms) L2-confirmation is O(sec) This L2-confirmation triggers two actions: - Action 1: the submission of a Tx to L1 to verify the 2/3 super-majority; this will be executed and confirmed within ~12mins (postconfirmation) - Action 2: certificate of
4K
franck cassez
@franckDownunder
Mar 15, 2025
Replying to @hrkrshnn
There is a difference between low level code like EVM bytecode and high level code like Solidity. IMHO there is some merit in having good high level abstractions (a DSL) to specify the business logics. That’s what’s Solidity lacks and what Move provides (and it borrows some
431
franck cassez
@franckDownunder
Feb 19, 2025
Replying to @franckDownunder and @SashaSpiegelman
it reminded me of the EF introducing EELS "a new standard for execution layer specification" after we (at ConsenSys) had already contributed the Dafny EVM github.com/Consensys/evm-…, a readable and executable EVM specification ... And there is even an EF Academic grant topic in
GitHub - Consensys/evm-dafny: An EVM interpreter in Dafny
From github.com
483